JSON AI Studio — AI assistant for JSON

jsonaistudio.com · an AI assistant for JSON: edit, fix and explain JSON in plain English, review every change as a side-by-side diff, and download valid JSON. Free, no signup — try it.

AI-powered configuration management platform — enable developers, product managers, and solution architects to create, modify, validate, and manage JSON configurations through natural language conversation.

Project Structure

json-ai-studio/
├── README.md                  ← This file: overview, quickstart, links
├── docs/
│     └── adr/                 ← Architecture Decision Records (all tech decisions)
│         ├── 0001-record-tech-decisions-with-adrs.md
│         ├── 0002-litellm-over-openai-sdk-for-llm-integration.md
│         ├── 0003-deepdiff-over-native-diff-for-backend-engine.md
│         ├── 0004-sse-streaming-over-full-json-or-hybrid-for-chat-responses.md
│         ├── 0005-few-shot-prompt-strategy-over-instruction-only-for-llm-diff-generation.md
│         ├── 0006-global-session-store-over-explicit-session-lifecycle-for-in-memory-storage.md
│         ├── 0007-fail-fast-error-handling-over-retry-pattern-for-llm-calls.md
│         ├── 0008-cashify-react-json-view-over-native-or-alternatives-for-tree-rendering.md
│         ├── 0009-sonner-over-react-hot-toast-or-native-for-toasts.md
│         └── 0010-lucide-react-over-native-or-alternatives-for-icon-library.md
├── apps/
│     ├── api/                 ← FastAPI backend (Python 3.12+)
│     │     └── pyproject.toml
│     │     └── src/json_ai_studio/main.py
│     │     └── ai/feature/mvp/requirements.md   ← Backend feature requirements
│     └── web/                 ← Next.js frontend (React + Tailwind)
│           ├── package.json
│           ├── tsconfig.json
│           ├── next.config.mjs
│           ├── src/app/layout.tsx
│           ├── src/app/page.tsx
│           └── ai/feature/mvp/requirements.md   ← Frontend feature requirements
└── (root)

Quickstart

API (Backend)

cd apps/api
uv sync                         # install dependencies
uv run uvicorn src.json_ai_studio.main:app --reload    # starts on :8000, /docs auto-generated

Web (Frontend)

cd apps/web
npm install && npm run dev      # starts on :3000

Configuration (environment variables)

Both apps read config from env files that are gitignored — copy the checked-in examples and fill in values:

cp apps/api/.env.example apps/api/.env          # backend
cp apps/web/.env.example apps/web/.env.local    # frontend

Backend — apps/api/.env

All variables are optional, but chat and explain need at least one LLM provider configured — set one (or more) of the keys below. Which providers/models are active is controlled by the deployment registry at apps/api/src/json_ai_studio/config/deployments.yaml (ADR-0016); a deployment whose key is missing or empty is disabled automatically at load time.

Variable	Purpose
GOOGLE_AI_STUDIO_API_KEY	Google AI Studio (Gemini/Gemma) key — create one at aistudio.google.com/apikey. Enables the google_ai_studio deployment.
OPENROUTER_API_KEY	OpenRouter key — create one at openrouter.ai/keys. Enables the openrouter deployment (free-tier models available).
NVIDIA_NIM_API_KEY	NVIDIA NIM key. Enables the nvidia_nim deployment (disabled by default in the registry).
OLLAMA_BASE_URL	Local/remote Ollama endpoint, e.g. http://localhost:11434. Unset disables the ollama_local deployment.
GENERAL_API_KEY	Shared anonymous quota-pool key (default dev-default-key).
DATABASE_URL	Async Postgres DSN, e.g. postgresql+asyncpg://json_ai:json_ai@localhost:5433/json_ai_studio. Unset → anonymous-only mode (login returns 503). Run migrations with uv run alembic upgrade head.
AUTH0_DOMAIN, AUTH0_AUDIENCE	Auth0 tenant domain (no scheme) + API identifier. Together with DATABASE_URL they enable optional login (ADR-0015).
ANON_*, USER_* quota knobs	Credit / token / request-rate limits for anonymous and logged-in users — defaults in settings.py.
LOG_LEVEL, LOG_DIR, LOG_*	Logging (ADR-0017). LOG_LEVEL=DEBUG logs full payloads (working JSON, prompts, LLM output, diffs).

See apps/api/.env.example for the full annotated list with defaults.

Frontend — apps/web/.env.local

Variable	Purpose
NEXT_PUBLIC_API_BASE	Backend API base URL. Dev: http://localhost:8000/api. Prod: /api (reverse-proxied).
AUTH0_DOMAIN, AUTH0_CLIENT_ID, AUTH0_CLIENT_SECRET	Auth0 Regular Web Application credentials (optional login).
AUTH0_SECRET	Session-cookie encryption secret — generate with openssl rand -hex 32.
AUTH0_AUDIENCE, AUTH0_SCOPE	Must match the backend's Auth0 API identifier, or tokens come back opaque and fail JWT validation.
APP_BASE_URL	Where this app runs; the Auth0 callback URL must be <APP_BASE_URL>/auth/callback.

See apps/web/.env.example for the annotated list.

Architecture Overview

┌──────────────────────┐   HTTP (SSE streaming)   ┌──────────────────────┐
│    apps/web          │ ──────────────────────►   │   apps/api           │
│    Next.js + React   │                           │   FastAPI + litellm  │
│                      │                           │                      │
│  • JSON Upload UI     │    POST /api/chat (SSE)   │  • LLM Integration   │
│  • Chat Workspace     │  ──────────────────────►  │  • Diff Engine       │
│  • Diff Viewer        │   thinking → diff × N     │  • Validation Engine │
│  • Version Sidebar    │   → complete              │  • Session Manager   │
│  • Export Buttons     │                           │                      │
└──────────────────────┘                           └──────────────────────┘

Documentation

Doc	Purpose
ADR README	All architectural decisions, one per file, numbered & dated
Backend Requirements	WHAT the backend API must implement (derived from PRD v1.0)
Frontend Requirements	WHAT the frontend UI must implement (derived from PRD v1.0)

License

Licensed under the Apache License, Version 2.0 — see LICENSE.